1 - The main reason is that your car RESEMBLES (not IS) a Faraday cage (even though, hey, we are talking about 10 meters - the smallest - wavelength here! The wave doesn't exactly "see" the car - it sees a material that is a mix of air, metal and silica). In the car, there are also electronics that COULD also produce noise... But it wouldn't be the main issue.
2 - Your car is NOT a Faraday cage. Even if shorter waves (GSM, UMTS, LTE, Wi-Fi, Bluetooth) DO "see" your car as a "shield", it is not a Perfect Electrical Plane. Of course, the radio waves get attenuated, but yet you have a signal. On top of that, silica (windows) is not metallic at all, so the waves "see" giant "holes".
3 - This is not quite true: you normally need to orient a DIRECTIVE antenna. The OMNIDIRECTIONAL ones radiate almost in the same way in every direction. Provided that the antennas in your cell phone are not ARRAY ANTENNAS, their radiation is quite similar to an omnidirectional one... Except that for obvious reasons, they won't radiate to the front of the phone (or your head). Please notice also that the current generation of mobile phones have a chipset that allows them to adjust power in order to maintain a stable connection... The meaning is: your phone always has a signal - but you can't see the power output that grants that (trust me, it changes).
4 - This is absolutely wrong. There are millions of kinds of antennas and the specs depend always on their purpose. Car antennas (smallest wavelength = 10m) are not 5m length. That is because nowadays we attach active elements to them as well as implement different shapes in comparison to a normal dipole (car antennas are in fact an excellent example). Finally, there is a wide range of PATCH-ANTENNAS whose radiation frequencies are mostly proportional to the length of their sides (but don't take this strictly as a law).
Source: Trust me, I am an engineer... And I work with antennas